54o 



NATURE 



[October 21, 1 922 



metric methods that had to be developed for this problem 

 involve a considerable amount of photometric, spectro- 

 scopic, and statistical detail when put on a quantitative 

 basis, and cannot be fully described in this article. 



The qualitative application of the photometric 

 methods, however, is simple. For example, we need 

 only assume that the brightest stars in a globular 

 cluster have the same actual luminosity as the brightest 

 stars in the solar neighbourhood, and we can readily 

 compute the distance necessary to give them the 

 apparent brightness that is measured. 



If we admit the similarity of globular clusters, it is 

 obvious that either the apparent magnitude or the 

 apparent diameter can give us the distances of them 

 all when once we have determined the distances of 

 those nearer the earth. In practice the distances of 

 the nearest clusters have been determined from studies 

 of their variable stars, of their blue stars (spectral 

 type B), and of their red giant stars ; and checked by 

 spectroscopically-determined absolute magnitudes and 

 by means of the relative diameters. All the methods 

 agree in giving distances of the same order of magnitude. 

 We thus find that the globular clusters range in distance 

 from seven thousand parsecs to values nearly ten times 

 as great. Their diameters are of the order of a hundred 

 parsecs. Their brightest stars are a thousand times 

 as bright as the sun. 



The Structure or the Galaxy. 



The result of most interest that comes out of this 

 photometric investigation is the enormous dimensions 

 of the super-system of globular clusters and of the 

 Galaxy. Once the positions in space are determined, 

 it becomes clear, as had already been suspected from 

 an inspection of the apparent distribution of clusters 

 in the sky, that globular clusters are a part of the 

 .Milky Way system. They are associated physically 

 with the system of stars, nebulae, and open clusters 

 which is more or less symmetrically arranged with 

 respect to the equatorial plane of the Galaxy. In 

 measuring the distances of the remotest globular 

 clusters, therefore, we are but measuring the depth 

 of our own galactic system. That the Milky Way 

 itself extends to distances as great as those indicated 

 by the clusters is shown by the presence within it of 

 highly luminous types of stars with apparent magnitude 

 15 and fainter. 



It has been known for many years that globular 

 clusters are not uniformly distributed in galactic 

 longitude. They are most numerous along the ed 

 of the southern Milky Way. That one-sided distribu- 

 tion is now recognised as an indication of the sun's 

 very eccentric position in the galactic system. In this 

 same southern part of the sky we find the densest 

 galactic star-clouds and the greatest frequent \ ol 

 faint novas and of other types of distant objects, 

 whii li is but further evidence ol the greater depth of 

 the galactic system in the direction of Sagittarius. 

 Also in that general direction are some obstructing 

 dark nebulae, which may be wholly responsible for a 

 peculiar phenomenon in the distribution of distan; 

 globular clusters, that is, in their seeming absence from 

 regions very close to (he galactic plane. If the ob- 

 strui ting material were removed, we might see, near 

 thi galactic plane, clouds of faint Milky Way stars 



NO. 2764, VOL. I IO] 



still more dazzling than those observed, and globular 

 clusters still more distant than those now known, and 

 hence find that the greatest diameter of the galactic 

 system is even larger than the value now assigned — 

 approximately 100,000 parsecs. 



The observable dynamical relations within and 

 without the Milky Way are suggestive of its origin. 

 No open clusters have yet been found outside the Milky 

 Way region, but hundreds are known within. North 

 and south of the galactic plane the globular clusters 

 are equal in number, and their distances from the plane 

 are much smaller than the greatest diameter of the 

 system. Their velocities, so far as now known, are 

 high. Many are approaching the galactic plane with 

 speeds that soon must bring them to it. Their present 

 positions and motions make orbital motion around the 

 Milky Way improbable. From the present evidence 

 as to mass, velocity, and distribution, there can be 

 little doubt but that the known globular clusters pass 

 to and fro through the star fields of the galactic system, 

 notwithstanding their observed avoidance, apparent 

 or temporary, at the present time. Every passage 

 must reduce the velocity and alter the form. The 

 hypothesis that these globular clusters are being 

 diverted by degrees into galactic regions, and gradually 

 robbed of their stars, is upheld by observation and is 

 not opposed by present dynamical theory. Although 

 we see few intermediates between tin/ globular and 

 the more typical open clusters, many of the character- 

 istics of the open groups strongly support the suggestion 

 that they are the remnants of globular clusters or of 

 other systems that have been assimilated by the 

 incomparably more massive galactic assemblage. 

 Nearly a dozen " moving " clusters, comprising thou- 

 sands of members, are recognised among the stars 

 within seven hundred light-years of the sun. 



Two important theoretical researches by Jeans are 

 of much significance in this view of galactic structure : 

 (i.) the form resulting from the interpenetration of two 

 clusters, 3 and (ii.) the necessity, in accounting for the 

 present orbits of long-period binaries, of assuming 

 their former existence in a much more compact stellar 

 field than now exists in the solar neighbourhood. 4 

 The high stellar frequency near the centre of a globular 

 cluster would certainly supply conditions favourable 

 for modifying orbits, and it also might aid in explaining 

 the origin of long-period binaries which is not other- 

 wise accounted for satisfactorily. 



The determination, with the aid of clusters, of dimen- 

 sions for the galactic system much larger than had been 

 clearly indicated by studies of the nearer galactic stars, 

 led lo a further examination of the stellar distribution 

 in the solar neighbourhood. The hypothesis that the 

 galactic system, as we now know it, has developed 

 from the combination of minor groups, suggests that 

 the brighter stars near the sun may to a large extent 

 be members of a local system that is imbedded in 

 and moving through the general star fields of the Milky 

 Way. This condition actually appears to be the case, 

 and hence the results on galactic dimensions, from 

 clusters and from the nearer stars, do not contradict. 

 Stars of spectral type B down to the sixth apparent 

 magnitude seem to be almost exclusively members of 

 a local cluster or cloud. Brighter stars of Class A are 

 also affiliated with the same system. 5 Probably all 



